Improving the accuracy of PSI-BLAST protein database searches with composition-based statistics and other refinements.

By Alejandro A. Schäffer, L. Aravind, Thomas L. Madden, Serge Shavirin, J. Spouge, Yuri I. Wolf, Eugene V. Koonin, Stephen F. Altschul. 2001

1,399 citations recorded by the source corpus.

Subfields: Molecular Biology, Spectroscopy

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

In Spectroscopy

Citations in the corpus

Hidden Markov model speed heuristic and iterative HMM search procedure (2010, methodology): “Iterative search approaches have been utilized for sequence to position-specific scoring matrix comparison methods, e.g. PSI-BLAST, as well as profile hidden Markov model to sequence scoring, e.g. SAM ’ s tar-get2k [3-6].”
Protein sequence design with a learned potential (2022, methodology): “Rosetta ab initio uses secondary structure probabilities obtained from Psipred 65 – 68 to generate a set of candidate backbone fragments at each amino-acid position in a protein.”
Database resources of the National Center for Biotechnology (2003, methodology): “This composition-based statistical treatment, used in conventional protein BLAST searches as well as PSI-BLAST (9) searches, tends to reduce the number of falsepositive database hits (11).”
Small but versatile: the extraordinary functional and structural diversity of the β-grasp fold (2007, methodology): “For queries and searches containing compositionally biased segments, the statistical correction option built into the BLAST program was used [124].”
A comparison of profile hidden Markov model procedures for remote homology detection. (2002, methodology): “This version included the improvements described in Schaefer et al. ( 21 ).”
Emergence of Fatal PRRSV Variants: Unparalleled Outbreaks of Atypical PRRS in China and Molecular Dissection of the Unique Hallmark (2007, methodology): “In order to clarify that the deletions exist only in the strains we isolated, we did BLAST searches in GenBank with both VR-2332 and HEN sequence segments flanking the deletions as queries [55].”
Protein-to-protein interactions: Technologies, databases, and algorithms (2010, methodology): “Starting from known interactions, the database is built on a two-step process: (i) determination of orthologs by using BLASTP [Schaffer et al. 2001]; and (ii) mapping an interaction if both the interactors have an ortholog into the human.”
Database resources of the National Center for Biotechnology Information (2006, methodology): “This composition-based statistical treatment, used in conventional protein BLAST searches as well as PSI-BLAST searches, tends to reduce the number of false-positive database hits (13).”
A Probabilistic Model of Local Sequence Alignment That Simplifies Statistical Significance Estimation (2008, methodology): “We have confirmed previous observations that even an ad hoc method to reduce the relative entropy per position (‘‘entropy-weighting’’; [43]) greatly improves search sensitivity in HMMER [27], although, puzzlingly, the same effect was not seen by PSI-BLAST’s authors [48].”
Protein sequence design with a learned potential (2020, background): “Rosetta AbInitio uses secondary structure probabilities obtained from Psipred [75, 76, 77, 78] to generate a set of candidate backbone fragments at each amino acid position in a protein.”
Log-odds sequence logos (2015, neutral): “Receiver-operating characteristic (ROC) analysis (Gribskov and Robinson, 1996; Sch€affer et al., 2001) is often useful for comparing retrieval methods.”
Protein database searches using compositionally adjusted substitution matrices (2005, methodology): “The first is the expert-curated “aravind103” data set [29], consisting of 103 query sequences, and associated true positive lists from a non-redundant version of the yeast (S.”